Bug 1903660

Summary:	Cannot install with Assisted Installer on top of IPv6 since network provider is not started
Product:	OpenShift Container Platform	Reporter:	Ori Amizur <oamizur>
Component:	Networking	Assignee:	Alexander Constantinescu <aconstan>
Networking sub component:	ovn-kubernetes	QA Contact:	Ori Amizur <oamizur>
Status:	CLOSED ERRATA	Docs Contact:
Severity:	high
Priority:	high	CC:	aconstan, anbhat, anusaxen, ercohen, rfreiman
Version:	4.6
Target Milestone:	---
Target Release:	4.7.0
Hardware:	All
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	No Doc Update
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2021-02-24 15:37:21 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Ori Amizur 2020-12-02 15:23:45 UTC

Description of problem:

Unable to install Openshift cluster above IPv6 with Assisted Installer since the bootstrap of OVNKubernetes plugin expects all the master nodes according to the replicas definition to be up before it starts the plugin.

The reason for the above is that in Assisted Installer there is no dedicated bootstrap node.  Instead, initially there are 2 master nodes and a bootstrap.  When the installation of the 2 master nodes completes, the bootstrap becomes the third master.

 Here is the place in the code that checks that the number of available masters is the same as the replicas count:

https://github.com/openshift/cluster-network-operator/blob/384ea2c5001c72e8870b1ac722e314ff1df5c2d5/pkg/network/ovn_kubernetes.go#L250

Version-Release number of selected component (if applicable):

Actual results:

The network provider is not started

Expected results:

The expected behavior is that if there is at least a single master node, the network plugin should be started.

Comment 1 Alexander Constantinescu 2020-12-02 16:06:16 UTC

Setting this to not-blocker as it's not a regression.

Comment 2 Eran Cohen 2020-12-02 16:07:18 UTC

This is how cluster-authentication-operator is stating without 3 master nodes:
https://github.com/openshift/cluster-authentication-operator/blob/release-4.6/pkg/controllers/readiness/unsupported_override.go#L58

This is what assisted-instalelr is doing to allow etcd to start with just 2 master nodes:
https://github.com/openshift/assisted-installer/blob/master/src/k8s_client/k8s_client.go#L152

This is what etcd dose for allowing the assisted-installer to start today:
https://github.com/openshift/cluster-etcd-operator/blob/e157623a52bf255524a2a4c0628281aa150dd225/pkg/operator/bootstrapteardown/bootstrap_teardown_controller.go#L150

This is the new implementation for scaling etcd:
https://github.com/openshift/cluster-etcd-operator/pull/449/files

Comment 3 Rom Freiman 2020-12-02 20:28:28 UTC

The question is if it's possible to do something for 4.7 timeframe with a very limited scope (protected by a flag that assisted installer will set).
For 4.8 it should be addressed in a correct way.

Comment 5 Alexander Constantinescu 2021-01-28 12:07:30 UTC

@Ori Amizur: can you coordinate with QE and validate that the assisted install now works?

Comment 6 Anurag saxena 2021-02-08 16:43:10 UTC

As per @oamizur , he has verified that the fix works fine on 4.7. Ori, pleae feel free to update verification steps here if needed

Comment 9 errata-xmlrpc 2021-02-24 15:37:21 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: OpenShift Container Platform 4.7.0 security, bug fix, and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2020:5633

Comment 10 Red Hat Bugzilla 2023-09-15 00:52:18 UTC

The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days